Materials Processing Reasoning on MatProcBench (Year+Type)
54.82AccuracyProvMind-Hybrid
Evaluation Results
| Method | Links | |
|---|---|---|
| ProvMind-HybridSetting=Hybrid2026.05 | 54.82 | |
| ProvMind-HybridSetting=Hybrid2026.05 | 52.84 | |
| Few-shotPrompting=Few-shot2026.05 | 49.33 | |
| Few-shotPrompting=Few-shot2026.05 | 48.95 | |
| RAGRetrieval=RAG2026.05 | 48.5 | |
| Zero-shotPrompting=Zero-shot2026.05 | 45.74 | |
| GraphRAGRetrieval=GraphRAG2026.05 | 45.7 | |
| Zero-shotPrompting=Zero-shot2026.05 | 45.47 | |
| GraphRAGRetrieval=GraphRAG2026.05 | 45.24 | |
| RAGRetrieval=RAG2026.05 | 43.32 |